Site‐directed mutagenesis

For Escherichia coli expression, pNIC28-Bsa4 plasmid harboring the MAPK1 or MAPK3 wild-type gene was used. The point mutations were introduced on wild‐type gene using QuikChange Lightning Site-Directed Mutagenesis (Agilent Technologies, Santa Clara, CA). Sequence analysis was performed to confirm the presence of the desired mutations and the absence of unwanted ones.

Protein expression and purification

N-terminally His-tagged MAPK1 and MAPK3 wild type and variants were expressed with phosphatase in E. coli cells Rosetta. E. coli cells were grown in LB medium containing kanamycin (30 µg/mL final concentration) at 37 °C until OD600 nm reached 0.6 AU. Protein expression was induced by adding 0.5 mM isopropyl‐β‐d‐thiogalactoside (Sigma‐Aldrich, St. Louis, MO), The culture was incubated overnight at 20 °C under shaking and centrifuged. The sedimented material was reconstituted in 40 mL of a solution consisting of 50 mM Hepes, 500 mM NaCl, 5 mM Imidazole, 5% glycerol (pH 7.5), referred to as the “Binding buffer.” This buffer also contained 0.5 mM Tris(2‐carboxyethyl) phosphine and a mixture of protease inhibitors without ethylenediaminetetraacetic acid (EDTA), which were obtained from Sigma‐Aldrich. The cells were subjected to sonication while kept on ice using a Vibracell 75,115 sonicator (SONICS, Newtown, CT) with a cycle of 3 s of sonication followed by 9 s of pause. After sonication, the lysate was subjected to centrifugation. Subsequently, the resulting supernatant was loaded into a Ni–NTA (Ni2 + nitrilotriacetate) affinity column pre-conditioned with Binding buffer. The column was sourced from GE Healthcare in Chicago, IL. The recombinant protein was eluted with 250 mM imidazole in Binding buffer, concentrated to a final volume of 2.5 mL (Amicon concentrator Ultra‐15, Millipore, Burlington, MA) and then applied to a PD‐10 prepacked column (GE Healthcare) to remove the imidazole. The hexahistidine tag was removed by incubation with His‐tag tobacco etch virus (TEV) protease overnight, at 4 °C. The mixture, which included the cleaved protein, the His-tag, and the TEV protease, was loaded into a Ni–NTA affinity column pre-conditioned with the Binding buffer. The protein without His‐tag was present in the flow through.

Protein concentration was determined using a molar absorptivity at 280 nm of 44,810 M−1 cm−1 referred to a 41.477 kDa molecular mass for MAPK1 wild type and variants, and a molar absorptivity of 43,320 M−1 cm−1 referred to a 43.222 kDa molecular mass, for MAPK3 wild type and variants. The protein purity was checked by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS‐PAGE) using precast NuPage 4–12% Bis‐Tris polyacrylamide gels (Thermo Fisher). Western blot analyses were performed on the wild type and the variants to assess the presence of the phosphorylated (P-MAPK1 and P-MAPK3) and non-phosphorylated (NP-MAPK1 and NP-MAPK3) form of the protein. The presence of both forms was confirmed through immunodetection using specific antibodies: the anti-ERK1/ERK2 (Thermo Fisher Cat. 44-654G) and the antibody directed against the doubly phosphorylated MAPK3 and MAPK1 (anti-phospho-ERK1/ERK2-Thr185, Tyr187-Thermo Fisher Cat. 44-680G). To activate both MAPK1 and MAPK3, including their wild-type and variant forms, the plasmid pGEX-KG-MEKR4F (an active mutant of MEK1 kindly provided by Prof. Melanie Cobb, Southwestern University, TX) was employed for co-expression of these proteins. The wild-type and variant forms of MAPK1 and MAPK3 were expressed as N-terminally His-tagged proteins within E. coli cells BL21(DE3)-pLysS and subsequently purified with the previously described method.

Spectroscopic measurements

The intrinsic fluorescence emission spectra (ranging from 300 to 450 nm) for both wild-type and variant forms of MAPK1 and MAPK3 were measured at a temperature of 20 °C. This was done using an LS50B spectrofluorometer (Perkin-Elmer) with protein concentrations in the range of 100–130 μg/mL. The samples were prepared in a solution containing 20 mM Tris/HCl at pH 7.5, 0.1 M NaCl and 200 μM DTT; a quartz cuvette with a path length of 1.0 cm was used.

Circular dichroism (CD) spectra for MAPK1 and MAPK3, both wild type and variants, were recorded in two regions. Far-UV CD spectra (covering the range of 190–250 nm) were obtained at protein concentrations between 100 and 130 μg/mL (equivalent to 0.13 AU at 280 nm). The protein samples were dissolved in 20 mM Tris/HCl at pH 7.5, 0.1 M NaCl and 200 μM DTT. A quartz cuvette with a path length of 0.1 cm was used.

Near-UV CD spectra (spanning 240–420 nm) were monitored with protein concentrations ranging from 1.0 to 1.3 mg/mL (equivalent to 1.3 AU at 280 nm). The buffer composition for these samples was 20 mM Tris/HCl at pH 7.5, 0.1 M NaCl, and 1.0 mM DTT, and a quartz cuvette with a path length of 1.0 cm was utilized. The CD measurements were taken using a Jasco-815 spectropolarimeter (Jasco, Easton, MD, USA), and the results were expressed as [Θ], representing the mean residue ellipticity. This calculation assumed a mean residue molecular mass of 110.

GdmCl-induced equilibrium unfolding

Wild-type and variant forms of MAPK1 and MAPK3 (final concentration of 75.0 μg/mL) were incubated at increasing concentrations of GdmCl (ranging from 0 to 8 M) at a temperature of 4 °C. This process was conducted in a solution consisting of 20 mM Tris/HCl at pH 7.5, 0.1 M NaCl and 0.2 mM DTT. After a 30-min incubation period to reach equilibrium, intrinsic fluorescence emission and far-UV CD spectra were collected at a temperature of 10 °C using a cuvette with a path length of 0.2 cm. To assess the reversibility of the unfolding, MAPK1 and MAPK3 wild type and variants at a final concentration of 0.75 mg/mL were unfolded in the presence of 8.0 M GdmCl, at a temperature of 10 °C, within a buffer consisting of 20 mM Tris/HCl at pH 7.5, 2 mM DTT and 0.1 M NaCl. After 5 min, the refolding process was initiated by diluting the unfolding mixture tenfold, at 4 °C, into solutions with progressively lower GdmCl concentrations. The unfolding experiments were performed in triplicate.

Thermal denaturation experiments

Thermal denaturation experiments were carried out by heating MAPK1 and MAPK3 wild type and mutants (100–130 μg/mL) in a 0.1-cm quartz cuvette, from 20 to 90 °C, in 20 mM Tris/HCl, pH 7.5, 0.2 mM DTT, 0.1 M NaCl at a heating rate of 1° × min−1 using a Jasco programmable Peltier element. The dichroic activity at 222 nm, along with the photomultiplier signal, was measured [2]. The contribution from the solvent for all thermal scans was monitored at increasing temperatures. Melting temperatures (Tm) were determined analyzing the first derivative of the ellipticity at 222 nm in relation to temperature, as explained in [27]. Each of these measurements was conducted in triplicate.

Enzyme activity assay and kinetic studies

The enzymatic activity of the purified P-MAPK1 and P-MAPK3 wild type and variants was measured by using Chelation-Enhanced Fluorescence (ChEF) method monitoring the incorporation of phosphate into a substrate peptide [41]. The activity of P-MAPK1 and P-MAPK3, wild-type and variant forms, was measured at a temperature of 30 °C using PhosphoSensR Peptide AQT0490 (provided by AssayQuant Technologies Inc., Marlboro, MA, USA) as the substrate, following the method described in [24]. The reaction mixture consisted of 50 mM Hepes at pH 7.5, 10 mM MgCl2, 0.1 M DTT, 0.012% Brij-35, 1% glycerol, 0.2 mg/mL BSA, 5.0 mM MgATP and AQT0490 peptide at concentrations ranging from 0.069 to 40 µM. The final volume of the reaction mixture was 0.4 mL. The reaction was initiated by adding various amounts (0.008–24 μg) of P-MAPK1 or P-MAPK3 wild-type and variant forms. These enzymes were diluted in a solution containing 20 mM Hepes at pH 7.5, 0.01% Brij-35, 0.1 mM EGTA, 5% glycerol, 1 mM DTT and 1 mg/mL BSA. The final enzyme concentrations ranged from 0.5 to 2381 nM for MAPK1 wild type and variants and from 8.0 to 90 nM for MAPK3 wild-type and variants. The increase in fluorescence intensity at 490 nm (with an excitation wavelength at 360 nm), which corresponds to the phosphorylation of the AQT0490 peptide by P-MAPK1 or P-MAPK3, was continuously monitored for a duration of 5 min using an LS50B spectrofluorometer (Perkin-Elmer). The kinetic data were subsequently analyzed using GraphPad Prism 7.03 (La Jolla, CA, USA). The reported results represent the mean values obtained from three separate experiments utilizing distinct enzyme preparations.

Temperature dependence of P-MAPK1 and P-MAPK3 activity

The temperature dependence of P-MAPK1 and P-MAPK3 catalytic activity was studied by measuring the enzyme activity as a function of temperature. The choice of substrate peptide concentrations for both wild-type and variant forms was based on their respective Km values and was consistently set below the Km value. To initiate the reaction, 2 μL of pure enzyme (maintained at 10 °C) was added with continuous stirring to a 0.4 mL assay mixture. This assay mixture contained 50 mM Hepes at pH 7.5, 10 mM MgCl2, 0.1 M DTT, 0.012% Brij-35, 1% glycerol, 0.2 mg/mL BSA, 5.0 mM MgATP and 1 µM AQT0490 (as explained in the enzyme activity assay section). The entire mixture was equilibrated at different temperatures (10, 15, 20, 25, 30, 35, 37, 40, 42 and 45 °C) within a temperature-controlled cuvette. The final enzyme concentration ranged from 0.5 to 1400 nM. The fluorescence intensity at 490 nm was continuously measured over 5 min. The activation energies (Ea) for the catalytic reaction were obtained by nonlinear fitting to the Arrhenius equation calculating the variation of enzyme activity as a function of temperature

where k (s−1) is the rate constant at temperature T (K), A is a reaction specific quantity, R the gas constant (1.987 cal × mol−1 × K−1), and Ea is the activation energy of the reaction.

Data analysis

The variations in intrinsic fluorescence emission spectra caused by GdmCl were measured in terms of the intensity-averaged emission wavelength, \(\overline}\), [36] calculated according to

$$\overline = \sum (Ii/\lambda i)/\sum (Ii)$$

(2)

where λi represents the emission wavelength, and Ii corresponds to the fluorescence intensity at that specific emission wavelength. The value \(\overline}\) is an integral measurement, which remains largely unaffected by noise and represents an indicator of alterations in the shape and position of the emission spectrum.

The equilibrium unfolding transitions induced by GdmCl, which were followed using changes in far-UV CD ellipticity or intrinsic fluorescence, were examined by fitting the data from both the baseline and the transition region to a two-state linear extrapolation model [38]. In detail, the data were fitted to the following equation:

$$\Delta G_}}} = \, \Delta G^}_ }}} + m\left[ }} \right] - }\,}\,\left( }}} } \right)$$

(3)

In this equation, ΔGunfolding represents the change in free energy associated with unfolding at a specific GdmCl concentration, \(\Delta G^}_} }}}\) is the change in free energy related to unfolding when GdmCl is absent, and m denotes the slope term, which quantifies how the unfolding constant (Kunfolding) changes with each unit increase in GdmCl concentration. R and T are the gas constant and the temperature, respectively, and Kunfolding is the equilibrium constant for the unfolding process. The model calculates the signal as a function of the GdmCl concentration:

$$y_ = \frac + s_ [X]_ + \left( + s_ [X]_ } \right)*\exp \left[ }_ }}} - m[X]_ } \right)/RT} \right]}}}_ }}} - m[X]_ } \right)}}} \right]}}$$

(4)

where yi represents the observed signal, yU and yN denote the baseline intercepts for the unfolded and native protein, sU and sN signify the baseline slopes for the unfolded and native proteins, [X]i indicates the GdmCl concentration after the ith addition, \(\Delta G^}_} }}}\) is the extrapolated unfolding free energy change in the absence of denaturant, and m represents the slope of a plot of ΔGunfolding versus [X].

The data were subjected to a global fitting procedure, with the m values shared among the datasets. No constraints were imposed on the other parameters. In accordance with Eq. (3), the denaturant concentration corresponding to the midpoint of the transition, denoted as [GdmCl]0.5, was determined as follows:

$$\left[ }} \right]_ = \Delta G^}_ }}} /}$$

(5)

GraphPad Prism 7.03 was used to fit all unfolding transition data.

Far-UV CD spectra obtained incrementing the GdmCl concentration were analyzed using a singular value decomposition algorithm (SVD). This analysis, performed by MATLAB software (Math-Works, South Natick, MA), aims to remove from the data the high-frequency noise and low-frequency random errors and to determine the number of independent components within a set of spectra, as outlined in [27].

For all the MAPK3 variants and certain MAPK1 variants, the GdmCl-induced equilibrium unfolding exhibited a non-two-state pattern due to the formation of an intermediate state at low denaturant concentrations. When an intermediate state was observed, the variations in [Θ]222 or intrinsic fluorescence resulting from the increasing GdmCl concentrations were fitted to the equation describing a three-state folding process:

$$F = \frac}\frac}} \right] - D50}}}} \right) \cdot \left( }\frac}} \right] - D50}}}} \right)} \right)}}}\frac}} \right] - D50}}}} \right) \cdot \left( }\frac}} \right] - D50}}}} \right)} \right)}}$$

(6)

where F is \(\overline}\) in Eq. (2), or [Θ]222, m represents the change in the solvent-accessible surface area involved in the transition. Specifically, D50I–N and mI–N denote the midpoint and m value for the transition from the native state (N) to the intermediate state (I), while D50U–I and mU–I represent the midpoint and m value for the transition from the intermediate state (I) to the unfolded state (U) [35]. FI which represents \(\overline}\) or the [Θ]222 at the intermediate state (I), is a constant. FN and FU, which represent the \(\overline}\) or the [Θ]222 of the N and the U state, respectively, show a linear dependence on GdmCl concentration:

$$F_}} = a_}} + b_}} \left[ }} \right]$$

(7)

$$F_}} = \, a_}} + \, b_}} \left[ }} \right]$$

(8)

In this equation, aN and aU represent the baseline intercepts for N and U, while bN and bU correspond to the baseline slopes for N and U, respectively. All the data related to unfolding transitions were fitted using GraphPad Prism 7.03.

Western blot

Following the SDS-PAGE procedure described earlier, the protein bands were transferred onto Hi-Bond N + membranes (Cytiva) using a XCell II Blot Module (Thermo Fisher) with a constant voltage setting of 30 V and 250 mA for 60 min. After blotting, the membranes underwent three washes with Tris-buffered saline (TBS) containing 0.1% Tween® 20 detergent (TBST). Subsequently, the membranes were blocked with TBST containing 5% Bovine Serum Albumin (BSA, Sigma-Aldrich). Following this, the membranes were washed once with TBST and twice with TBS before being exposed to the primary Antibody 44-680G, which is an anti-phospho-ERK1/ERK2 (T185, T187) antibody from Thermo Fisher, raised against the doubly phosphorylated ERK2. The primary antibody was diluted to a 1:5000 concentration and incubated overnight at 25 °C.

Afterward, the membranes, which now had the primary antibody bound to them, were incubated with anti-Rabbit IgG (Thermo Fisher, Catalog # 32,460), diluted to a 1:5.000 concentration in TBS, for 60 min at 25 °C. They were then subjected to three additional washes in TBST. Labeled protein bands were detected using a chemiluminescent system according to the supplier's protocol (Amersham).

The same western blot membranes were subsequently used for re-probing with an antibody specific to the non-phosphorylated form of ERK2, labeled as antibody anti-ERK1/ERK2 (Thermo Fisher, Catalog# 44-654G), after implementing the stripping protocol outlined in [24].

Molecular dynamics

The crystal structures of the wild-type MAPK1 and MAPK3 were acquired from the Protein Data Bank, representing both the phosphorylated/active state (P-MAPK1: 5v60, P-MAPK3: 2zoq) and the inactive/non-phosphorylated state (MAPK1: 4zzn, MAPK3: 4qtb). To investigate the impact of specific mutations, we generated nine variants for both NP-MAPK1 and P-MAPK1 using in-silico point mutation with PyMOL (The PyMOL Molecular Graphics System, Version 1.2r3pre, Schrödinger, LLC). The variants included E33Q, E81K, R135K, D162G, D321G, D321N, E322K and E322V. Additionally, we examined a variant of MAPK3, namely E98K, which corresponds to the E81K variant observed in MAPK1.

To investigate the dynamic behavior of all twenty-four distinct structures, including both the wild-type proteins and the variants (both phosphorylated and non-phosphorylated), classical molecular dynamics (MD) simulations were conducted using the GROMACS package [11, 14] with the GROMOS43A1-P force field [21]. The MD simulations allowed us to observe and analyze the conformational changes and interactions within the protein structures over time, providing valuable insights into their functional characteristics and stability.

MD simulations were carried out in the NPT ensemble at 315 K and neutral pH by following the here reported procedure. The temperature is held fixed at 315 K using the v-rescale thermostat [4] with a coupling time of 0.1 ps. The pressure is kept constant at the reference pressure of 1 bar with a coupling time of 1 ps and an isothermal compressibility of 4.5·10−5 bar−1, exploiting the features of the Parrinello-Rahman barostat [26]. The simple point charge model is used for water molecules. The simulation box is cubic (with a side of 9.53 nm), and each protein is immersed in an appropriate number of water molecules (in order to have the liquid water density), and Na+ and Cl− counterions at a concentration of 0.15 M are added to have a wholly neutral system. Periodic boundary conditions are used throughout the simulation, and the particle mesh Ewald algorithm is used to deal with the long-range Coulomb interactions [6]. A time step of 2 fs was used. A non-bond pair list cutoff of 1.0 nm was used, and the pair list was updated every ten steps. Each protein structure is initially relaxed in vacuum via a steepest descent minimization. Then, the appropriate amount of counterions and water is added. At this point, the whole system, solute and solvent, is relaxed again via a steepest descent minimization and then equilibrated for 100 ps in the NVT ensemble at three increasing temperatures: 150 K, 293 K and 315 K. This is the situation from which the final 800-ns-long NPT MD simulation (for each system) is started.